Shifting the Hydrolysis Equilibrium of Substrate Loaded Acyl Carrier Proteins

Abstract

Acyl carrier proteins (ACP)s transport intermediates through many primary and secondary metabolic pathways. Studying the effect of substrate identity on ACP structure has been hindered by the lability of the thioester bond that attaches acyl substrates to the 4’-phosphopantetheine cofactor of ACP. Here we show that an acyl acyl-carrier protein synthetase (AasS) can be used in real time to shift the hydrolysis equilibrium towards favoring acyl-ACP during solution NMR spectroscopy. Only 0.005 molar equivalents of AasS enables one week of stability to palmitoyl-AcpP from Escherichia coli. 2D NMR spectra enabled with this method revealed that the tethered palmitic acid perturbs nearly every secondary structural region of AcpP. This technique will allow previously unachievable structural studies of unstable acyl-ACP species, contributing to the understanding of these complex biosynthetic pathways.

Graphical Abstract

Acyl carrier proteins (ACP)s play a critical role by chaperoning intermediates between enzymes in many primary and secondary metabolic pathways.^1,2,3 Upon binding to the appropriate partner enzyme, the 4’-phosphopantetheine (PPant) “arm” of ACP facilitates exiting of the acyl substrate from within the ACP pocket and its entering into the active site of a partner enzyme in a process termed chain-flipping.⁴ Many enzymes demonstrate specificity for particular acyl chain lengths or oxidation states. Since the acyl substrate is sequestered inside the hydrophobic pocket of ACP during partner enzyme binding,⁵ protein-protein interactions must play a critical role in the processivity of these pathways.⁶ However, little is understood about how the identity of the substrate is communicated by the ACP or how substrate identity effects the overall structure of loaded ACP. We have sought to understand these processes using solution phase NMR spectroscopy.

Acyl substrates are covalently attached to the terminal thiol of PPant through a thioester linkage. Sequestration of hydrophobic acyl chains inside the pocket of ACP is known to protect the labile thioester bond from hydrolysis, and longer chains have been shown to have increasingly greater hydrolysis rates,⁷ a phenomenon that has impeded the study of long chain acyl-ACPs. Chemoenzymatic methods have been devised to replace the thioester with a more stable amide bond, however this has been shown to affect the structure of ACP.⁸ Therefore, we sought to develop a method that would allow structural study of long chain acyl-ACP. Recently, a method was described to probe molecular influences of labile substrates in carrier proteins from non-ribosomal peptide synthases (NRPS),⁹ which allowed the first solution NMR structure of a substrate loaded peptidyl-carrier protein.¹⁰ We realized that a similar strategy could be employed by leveraging the acyl acyl carrier protein synthetase (AasS) to shift the hydrolysis equilibrium towards favoring acyl-ACP during NMR experiments.

AasS is a promiscuous synthetase that attaches fatty acids to holo-ACP, forming acyl-ACP and consuming ATP.¹¹ This reaction opposes hydrolysis, therefore we hypothesized that AasS and reaction components could be combined with purified acyl-ACP as a tool to re-attach fatty acids to hydrolyzed holo-ACP (Scheme 1), thereby preserving the acylated form for long protein NMR experiments. As described herein, we have evaluated the stability of short, medium, and long acyl chain lengths at timescales relevant to solution NMR structural studies. We have also determined the optimal AasS concentration with which to collect NMR spectra. This method allowed the collection of 2D NMR spectra of palmitoyl-AcpP, a species which would have not otherwise been obtainable, leading to the discovery that addition of this long chain fatty acid perturbs nearly every region of AcpP secondary structure. This concept can be further applied to uncover the effect of other labile intermediates upon the structure of ACP and its partner enzymes to further understand communication and processivity in these complex, carrier protein dependent, biosynthetic pathways.

Scheme 1.

Shifting equilibrium towards acyl-ACP workflow. (A) ACP overexpression typically yields two species. (B) apo-ACP, without PPant cofactor, and holo-ACP, with PPant covalently bound. A PPTase attaches CoA to apo-ACP, producing holo-ACP to which AasS attachs a fatty acid. (C) Purified acyl-ACP added to NMR tube along with AasS and reaction components to counteract hydrolysis by re-attaching fatty acids to holo-ACP.

Using ACP from Escherichia coli fatty acid synthase (AcpP) we first tested the stability of acyl-AcpP in the presence or absence of 0.005 molar equivalents of recombinant Vibrio harveyi AasS¹² in buffer conditions amenable to solution phase NMR. Using conformationally sensitive urea-polyacrylamide gel electrophoresis (PAGE), the acylation state of AcpP was monitored over the course of a week, an amount of time typically required to perform triple resonance NMR experiments and analyze protein structure.

Short-chain butyryl- (C4-), medium chain octanoyl (C8-), and long-chain palmitoyl- (C16-) fatty acids were evaluated, each showing significantly different stability profiles. C16-AcpP showed a dramatic enhancement by AasS. Without AasS, holo-AcpP and dimers of holo-AcpP immediately form, and their levels increase relative to C16-AcpP with each subsequent day. Excitingly, with AasS present, an entire week of C16-AcpP is afforded. C8-AcpP was expected to be the most stable, with the thioester protected by sequestration.^5,13,14 Accordingly, C8-AcpP was extremely stable in both in the absence and presence of AasS for a week. C4-AcpP, which runs as multiple bands in urea-PAGE indicative of multiple sequestered states,¹⁵ is also retained through AasS activity. In the absence of AasS, bands corresponding to holo-AcpP and holo-AcpP dimers become clear after just one day, though about half of the population remains in the form of acyl-AcpP for an entire week (Figure 1).

Figure 1.

Acyl-AcpP stability time course of long, medium, and short chain lengths. C16-AcpP, C8-AcpP, and C4-AcpP are shown in (A), (B), and (C) respectively. Purified acyl-ACP and reaction components were monitored for a week in the presence or absence of AasS with 5 mM ATP, 5 mM MgCl₂, and 1 mM fatty acid in 50 mM phosphate buffer pH 7.4 by urea-PAGE.

For NMR experiments, AcpP was expressed with ¹⁵N isotope enrichment, while AasS was expressed with natural abundance and added to the NMR sample. Therefore, only the chemical shift of AcpP residues was detected as AasS served to counter the hydrolysis of acyl-AcpP to holo-AcpP. To first determine whether AasS would perturb the NMR spectra, we titrated increasing molar equivalents of AasS to the stable ¹⁵N-C8-AcpP species, acquiring ¹H-¹⁵N HSQC spectra at each titration point. The presence of AasS did not perturb the structure of AcpP, however, significant signal loss preventing peak resolution was observed above AasS molar equivalents of 0.025 (Figure S1).

With an increase in stable acyl-AcpP, we sought to gain structural information about each species using solution NMR. ¹H-¹⁵N HSQC spectra of holo-AcpP and C16-AcpP in the presence and absence of AasS were obtained over the course of 1 hour and compared. Without AasS present, the spectra overlay exactly, indicating complete hydrolysis of the acylated species (Figure S2). In the presence of AasS, the C16-AcpP spectrum shown in Figure 2 was obtained. This data revealed that nearly every secondary structure region of AcpP is perturbed by addition of a palmitoyl group, including loop I, helix II, loop II, helix III, and helix IV (Figure 2).

Figure 2.

(A) Overlaid ¹H-¹⁵N HSQC spectra of holo-AcpP and C16-AcpP in 50 mM phosphate buffer, pH 7.4. (B) Zoomed in view of selected peaks with large chemical shift differences. (C) Chemical shift perturbation (CSP) plot comparing holo-AcpP to C16-AcpP. The solid line represents the mean CSP and the dashed line represents one standard deviation above the mean. Those CSP values greater than one standard deviation of the mean are colored in magenta. (D) Residues with CSPs larger than one standard deviation of the mean shown as sticks on AcpP crystal structure PDB ID: 2FAD and colored magenta.

The thioester bond attaching substrates to ACP is susceptible to hydrolysis, with longer acyl chains hydrolyzing more readily.⁷ Molecular dynamics simulations have shown that the hydrophobic pocket of AcpP expands to accommodate growing acyl substrates up to 10 carbons in length before the acyl chains can no longer fit in the pocket.¹⁶ This explains why longer chain lengths are less sequestered, leaving the thioester more readily susceptible to hydrolysis. Replacing the sulfur with more stable atoms has been used to stabilize this bond for NMR data collection, but these have been shown to perturb ACP structure, emphasizing that ACP is sensitive to even one atom modification of substrates.⁸ The effect of substrate identity on ACP structure should provide further insight into the communication mechanisms which allow ACP to distinguish between over 25 enzymes in E. coli^17,18,19 and deliver substrates to the appropriate partner enzyme. The method described here to stabilize acyl-ACP by shifting the hydrolysis equilibrium through the use of AasS will allow for structural studies of substrate loaded ACP that were previously inaccessible.